ROVing the Depths

Were it possible to take a wide-angle photograph of the Gulf of Mexico’s busiest oil fields a mile underwater, the image would bring to mind an otherworldly city scene from a science fiction novel: clusters of pipe-sprouting wellheads, thick pipelines in trenches, caisson-shaped “subsea separation” stations that separate gas from oil and water, giant tube-like suction anchors, and electrically-powered pumping stations on steel frames. Dozens of boxy underwater craft, known as remotely operated vehicles (ROVs), move silently through this strange world, turning valves, attaching hoist cables, connecting cables, and laying pipe.

Although this environment is found on Earth, it as hostile to humans as interstellar space, with its near-freezing temperatures and immense pressures of 2,400 psi, enough to crush an unprotected person. While manned submersibles with manipulators can handle such pressure, they are cumbersome and inefficient compared with the sturdy, hermit crab–shaped ROVs. As a result, these underwater cities are mostly installed and maintained remotely by ROVs.

Deep-drilling offshore operations from West Africa to Asia, along with businesses that install underwater power and communication cables, would be dead in the water without “work class” ROVs. Not only essential to drilling and production, they are key to emergency operations such as helping to fish out giant pieces of drilling equipment that have fallen from the surface. Such ROVs are typically about the size of a minivan and cost a few million dollars apiece. Other types range in size from recreational, camera-toting models of a few dozen pounds to a 60-ton leviathan that gouges pipeline trenches with blasts of high-pressure water. None can accommodate people. IQ-wise, the carpet-sweeping Roomba outthinks them all, because the long extension cord behind every ROV brings not just electrical power but orders from above. (Technically, those reporting on the Deepwater Horizon oil spill are incorrect when they refer to ROVs as robots. Any robot worth its name is off leash and operates with some element of artificial intelligence.)

The job of ROVs has come into focus during the Deepwater Horizon disaster, during which BP marshalled more than a dozen ROVs for what the industry calls “subsea intervention.” The attempt to stop the oil streaming from the well could not occur without these multi-function machines. Topside, operators comfortably reclined in padded chairs while they stared at a bank of monitors with video feeds from cameras attached to the ROV. Each operator manipulated a joystick on his left armrest to move the ROVs’ mechanical arm; the joystick on the right armrest controlled the motion of the vehicle through the water. The most popular ROV in use during the spill was Oceaneering’s Millennium, an 11.5-foot-long, 8,000-pound, foam-topped, rectangular machine.

The ROVs aided the engineers fighting to shut off the well hemoraghing millions of gallons of oil by using their cameras and instruments to monitor the status of critical hardware, such as the riser, wellhead, and blowout preventer; making connections, such as putting hoses onto fittings; opening valves; and repositioning chunks of heavy iron being lowered on cables from ships floating above. 

In the early underwater era, any work at depth was done by divers wearing standard rubber, canvas, and metal suits. Rarely venturing beyond 200 feet or so, divers bolted together pipelines, patched ships, and cut treasure from shipwrecks, clomping along the seafloor in weighted boots and canvas suits, supplied with air pumped into brass helmets through rubber hoses.

Advances in scuba tanks and regulators, decompression equipment, and specialized deep-diving mixtures of hydrogen, helium, and oxygen freed commercial divers from hoses and tenders, making it possible to work much deeper. Divers could reach 2,200 feet, but at enormous risk. Experts at the time concluded that the practical limit of mixed-gas work dives was 1,000 feet. Even at this shallower level, costs per man-hour remained extremely high. Almost any mishap to a diver could be life-threatening, and a rescue would shut down work on the entire site.

At first the hazards and limits of deep diving seemed to point toward the manned submersible, a minisubmarine from which a small crew could wield tools with manipulator arms. In the mid-1960s high-tech corporations practically swooned over such machines, because they would open up the deep to mineral exploration. The mania for manganese nodules and other bounty touched even companies with no experience in the subject. American corporations investing in undersea tech included Westinghouse, Northrop Grumman, North American Rockwell, General Motors, Reynolds Metals, Lockheed, General Dynamics, Hughes Tool, and Litton Industries. “Every major defense contractor went into it,” recalls Robert Wernli, a retired ROV engineer for the Navy who is now a consultant. “But after they found the certification requirements were so expensive to meet, most of the machines built went on blocks for display.” Even General Mills, maker of cake mixes and cereals, got into the act by winning a government contract to build the deep-diving Alvin, later turning over the work to Litton.

“Most of the companies entered because they saw Howard Hughes getting involved, but he was after a Russian submarine instead,” explains Drew Michel, owner of ROV Technologies Inc., referring to the CIA-backed USNS Glomar Explorer, a deep-sea drillship built to recover the Russian K-129 submarine that sank in the Pacific in 1968. One of the few competing efforts to hit the water was Beaver Mark IV, built by North American Rockwell and equipped with manipulators to install oil and gas equipment in waters as much as a half-mile deep—an underwater workboat that carried the nickname “Roughneck.”

Assuming that oil field equipment would always need the human touch, most concepts for deepwater oil fields by 1973 assumed that fleets of diving bells and submersibles would shuttle workers from the surface down to steel capsules that would form a compact but habitable refuge enabling “shirtsleeve” access to wellheads and pumps. Workers equipped with breathing masks to protect them from asphyxiating gases would climb from submersible to capsule, perform their jobs, and then return to the surface. Through the 1970s oil companies conducted experiments with manned submersibles and subsea work enclosures.

Until 1970 ROVs had been trapped in a niche market, serving a small number of naval and scientific customers. Aside from some early military trials—the U.S. Navy lowered an undersea TV camera to check on A-bomb blast damage to shipwrecks near Bikini Atoll in 1947, and the British Navy used another to confirm the identity of the sunken submarine Affray in 1951—credit for the first working ROV probably goes to French underwater archaeologist and photographer Dimitri Rebikoff for his work in the early 1950s. Frustrated by the fact that some Mediterranean wrecks were too deep for divers to investigate, he installed a camera in a pressure-resistant housing, added a water-correcting lens, and mounted it on a tethered vehicle that he dubbed Poodle. To increase its treasure-finding capacities, he added a magnetometer and sonar set. (Most ROV experts credit Poodle as the world’s first ROV. Having recently built a diver-driven, one-man underwater scooter called Pegasus for use by the Submarine Alpine Club of Cannes, France, Rebikoff had a head start in building Poodle’s controls and power train.) On its first dive in 1954, Poodle sent up video footage of two unexplored Phoenician wrecks, one 700 feet down.

In the next two decades, U.S. Navy labs and contractors built more camera-carrying ROVs, one of which, Snoopy, was notable for its reliance on direct hydraulic drive, transferred from tender to vehicle through a long tether hose. (Today’s tethered ROVs all rely on electrical power, as did the successor, Electric Snoopy.)

            The first ROV capable of substantial work—the Cable-Controlled Underwater Recovery Vehicle (CURV)—emerged from a Navy laboratory in Pasadena, California, in the mid-1960s. Originally tasked with recovering torpedoes that failed to surface after test shots, CURV-I made international news in 1966 off Palomares, Spain, where an H-bomb had plummeted into the Mediterranean after the B-52 carrying it collided with a KC-135 tanker during midair refueling. The bomb came to rest precariously on the lip of a steep slope in a skein of parachute shrouds. A rescue mission by the manned submersible Alvin nearly ended in disaster when it got tangled in the cords while trying to attach lifting shackles. Alvin wiggled free and the Navy sent in CURV-1, even though the bomb lay at 2,850 feet, well below the ROV’s rated depth. It reached depth without imploding and successfully completed the rigging job. Seven years later CURV III helped rescue the two-man crew of a submersible stuck on the seafloor off the Irish coast.

The Navy’s Remote Unmanned Work System, deployed after CURV, was directed at search and recovery work. “The idea was to go to 20,000 feet with all the tools you needed to recover a black box,” says Wernli. Such a depth range would enable this breed of ROV to access nearly all objects on the ocean floor. Cable handling proved an immediate challenge: if connected directly to a boat on the surface via a single, thick cable more than four miles long, the ROV would have been at the mercy of any deep currents straining on the line. A solution emerged that involved using two cables: a strong umbilical line reinforced with Kevlar fiber that plummeted straight down from the tender to a base station—known as the Primary Cable Termination—that hung just above the seafloor; and a lightweight and neutrally buoyant tether that was paid out horizontally as the ROV ventured off to work nearby. This arrangement kept cables from dragging along the seafloor and stirring up a cloud of talcum-fine silt that would block visibility for hours. (Work-class ROVs use the same arrangement today. Now the umbilical terminates at a strong metal cage that serves as a garage for the ROV when not in use.)

Good military R & D notwithstanding, a reliable and functional fleet of ROVs had to wait for commercial demand. This eventually arose from the peculiar requirements of the North Sea offshore oil fields, where exploratory wells revealed large reserves of oil and gas. The economics looked even better after the first oil crisis boosted crude prices. Production began in 1975. The demands of the setting—undersea wellheads that had to be connected to pipelines far from shore, work dogged by frequent storms—were so novel that, by 1980, development and installation costs had exceeded NASA’s budget for the the Apollo program. Reserves estimated at 70 billion barrels made them worthwhile.

The early years saw dozens of manned submersibles and hundreds of divers at work, with small numbers of ROVs of uneven reliability at the margin. Keeping high-pressure seawater out of electrical equipment remained a large challenge. One particularly useful ROV was the “flying eyeball” model, which trained its camera on a diver to monitor his safety. By 1980, as abilities expanded and reliability improved, the North Sea ROV population surged. (The far-seeing award should go to Shell Oil Co., which began developing an underwater wellhead system in 1958 that depended on an ROV-like machine called UNUMO, short for “universal underwater MOBOT.” UNUMO, which was tested in the early 1960s off California, traveled around a wellhead on a steel track while performing its maintenance tasks.)

Although the North Sea fields lie in shallower waters than those off West Africa, Brazil, and in the Gulf of Mexico, that market paved the way for critical deepwater advances: connectors that didn’t short out in seawater; acoustic beacon navigation; robust manipulators; color video for use when the water was clear; and sonar for when it was not.

Changes in underwater hardware were equally important. “That was the big turning point,” recalls Wernli. “When [oilfield engineers] accepted they had to go deep, they started designing the equipment for operating subsea equipment and turning on and off valves. The key was that there had to be standard docking so the ROV could plug something in or manipulate something.” For example, if a shaft on a valve assembly needs turning, it should have a handle designed specifically to be gripped by high-torque, rotating claws. (ROVs don’t need the leverage that humans get from a big crescent wrench.) Today operators can preprogram their ROVs to park at a designated spot, anchor themselves by taking a grip on a stout piece of steel framework, and reach out to adjust a set of valves; this reduces the risk of error and speeds the work.

On the Deepwater Horizon “spill cam” Web sites, viewers have seen ROVs handling shears, circular saws, and diamond-wire cutters. ROVs can also carry drills, abrasive wheels, and jets to cut steel with high-pressure water and abrasive powder. Such a toolbox will be handy for deepwater decommissioning work, that costly phase at the end of a well’s useful life when oil companies are obligated by federal regulations to cut away old pipes, valves, and other sea-bottom steelwork and hoist it to the surface. Nothing is to be left showing above the mudline, except where structures can serve as artificial reefs. Decades from now, deep demolition may be the job of heavy work-class autonomous underwater vehicles (AUVs), which will fetch tools and recharge from unmanned workboats floating nearby. Cost-conscious oil companies also hope that AUVs can take over the routine maintenance of subsea equipment from ROVs. This will allow labor-intensive ROVs to focus on expert jobs such as “workovers” of aging wells, rooting out clogs, and replacing corroded parts.

Meanwhile AUVs head out on relatively simple missions, finding their own way back home with inertial guidance and GPS fixes from the surface. Short-range acoustic beacons permit precise navigation through built-up areas. AUVs already excel at sea-bottom sonar mapping, oceanographic research, and mine hunting for the Navy. If given the opportunity to recharge at intervals, AUVs can work for days or weeks before returning to base.

Many in the deepwater oil industry compare their technological prowess to the glory days of Apollo. Therein lies a lesson, drawn from many trials of deep-diving humans vs. machines since 1970, which may be troubling to advocates of manned space missions. Strictly on a business basis, humans can’t compete with remotely controlled machines when it comes to doing a large volume of high-stakes work under distant and dangerous conditions. Today’s ROVs still depend on human operators who watch the action through surrogate eyes and pull the strings from a comfortable distance. But that era will pass, as artificial intelligence linked to machine vision advances on a fast and irreversible track. With each passing year, claw-waving machines, already populous underwater and one day swarming into space, will advance in skill and judgment, executing ever more complex missions.